Live Translation
Translate speech in real time, one-way into a single language or two-way between two languages.
Kalima can translate speech as it happens, showing the translated text right next to the original transcript. Use it to caption a presenter for a multilingual audience, follow a cross-language conversation, or produce live-translated subtitles, all without stopping the session. You can translate between more than 60 languages.
What live translation does
When translation is on, Kalima transcribes the spoken audio as usual and produces a translation alongside it in real time. You see both versions together: the original words in the language they were spoken, and the translation inline with them. The translation keeps pace with the live transcript, so everyone in the room can follow along in the language they understand.
There are two modes, and you choose one when you set up the session. One-way translation renders everything spoken into a single target language you choose. Two-way translation handles two languages in both directions at the same time, which suits a back-and-forth conversation.
Choosing one-way or two-way
One-way translation takes everything that's spoken and translates it into a single target language. This is the right choice when one or more people speak, and the audience needs it all in one shared language.
Good for:
- A presenter speaking to a multilingual audience
- Captioning a talk or lecture into the audience's language
- Translating a video or call into your own language
Turn on translation when you start a session
You set up translation before you start recording, as part of the session setup. It only takes a moment.
The languages you pick for translation are managed automatically as part of your session's expected languages, so Kalima already knows what to listen for. You can still add other expected-language hints if more languages than the translation pair will be spoken.
Reading the original and the translation together
Translated text is shown inline with the live transcript, so you can read the original speech and its translation side by side as the session runs. The same view carries through to the Studio transcript editor after you stop, where you can switch between viewing the original only, the translation only, or both at once.
Each part of the transcript also carries a language badge, so in a multilingual session it's clear which language was spoken before it was translated. Learn more about how Kalima labels languages on the Languages and Detection page.
Translation quality
Translation is available on every plan. On higher plans, Kalima uses an enhanced AI translation that produces more natural, context-aware results. This is included automatically on Pro, Team, and Enterprise. On Free and Plus, you get standard real-time translation.
| Plan | Translation | Quality |
|---|---|---|
| Free | Included | Standard |
| Plus | Included | Standard |
| Pro | Included | Enhanced |
| Team | Included | Enhanced |
| Enterprise | Included | Enhanced |
If translation accuracy matters for specialized content such as product names, medical or legal terms, or people's names, add session context before you record. You can even supply preferred translations for specific terms so they come out the way you want.
Sharing and exporting translations
A translated session is just as useful afterward as it is live.
- Live viewers of a shared session see the translation update in real time alongside the transcript, which is great for conferences, classrooms, and accessibility. See Live Viewer Mode.
- When you export your transcript as subtitles (SRT or VTT), you can choose whether the caption track contains the original text or the translation, so you can hand off captions in the language your audience needs.
Best practices
Add your translation languages and any expected-language hints before you start. Translation tracks the live transcript, so the more accurately Kalima recognizes the spoken language, the better the translation reads. For two-person conversations, two-way mode keeps both sides clear; for one speaker to many listeners, one-way into the audience's language is simplest.
Frequently asked questions
Yes. In the Translation step choose One-way (everything into one target language) or Two-way (between two languages in both directions). Translated text appears inline with the live transcript as you speak.
One-way translates all speech into a single target language you pick, which is ideal when one person speaks to an audience. Two-way translates between two languages in both directions at the same time, which is ideal for a conversation where each person speaks a different language.
No. Kalima detects the spoken language automatically and labels each part of the transcript. The languages you choose for translation are added to your session's expected languages so recognition stays sharp. You can add more expected-language hints if other languages will be spoken too.
Translation is included on every plan. Pro, Team, and Enterprise automatically use enhanced AI translation for more natural, context-aware results. Free and Plus use standard real-time translation.
Yes. When you share a live session, viewers see both the transcript and its translation update in real time. You can read more on the Live Viewer Mode page.
Yes. When you export subtitles (SRT or VTT) for a session that has translations, you can pick whether the caption track holds the original text or the translation. See Exporting Transcripts.